niftools/niflib

Niflib - Niftools C++ library for Nif IO

niflib's Issues

NifDoc v0.9 support

The following library does not support Nif.xml 0.9, making it impossible to use with newer nif programs.

Is there any plan for updating this repository and/or the docsys? (I was forced to use a fork of it since the normal one does not generate the files with the v0.9)

NiBoneLODController

Opens up access to the shape-collection.

https://www.dropbox.com/s/538k841388hopt9/NiBoneLODController.cpp.diff?dl=0
https://www.dropbox.com/s/f9uk6zpyzpjoi3j/NiBoneLODController.h.diff?dl=0

This is a good pattern - and probably can be used at various places. Would it make sense to factor this out into utility functions say:
template T bool GroupAddElement( vector<Ref > group, Ref elem)
template T bool GroupRemoveElement( vector<Ref > group, Ref elem)
template T bool GroupReplaceElement( vector<Ref > group, Ref old, Ref new)

Yes, I pushed a few diffs with these group-related interfaces. I probably have even some more just here in the code. I use it extensively to substitute nodes by new ones in NIFopt.
You still need to add it to each class though, as the group is private.

Of course. We also need to pass the count so it can be updated.

NiSkinData complement

Files require review

Allows access to the Bone-weights.

https://www.dropbox.com/s/s1xk3cq0y281o9w/NiSkinData.cpp.diff?dl=0
https://www.dropbox.com/s/od2qjuiesbz9707/NiSkinData.h.diff?dl=0

My feeling is that this patch goes against niflib's current design - it feels out of place. We have Get and Set functions for everything else elsewhere.
Could you rewrite it to fit that design?

(The other options is of course simply to make all data public... I'm actually not at all against that, but it would be a completely different library, and a lot of things would need to be readdressed.)

I can't really follow. It is a getter and a setter, and one that is in exactly the same way everywhere, just take GetVertices()/SetVertices(), it's almost a copy'n'paste job here only it's for bone-data. :^)

I don't actually care all that much, but:

1. Your getter and setter are not symmetric (getter = bone data, setter = vector of weights).
2. Your getter returns a raw pointer to the data; in other parts of the lib, a copy of the data is returned (IIRC).

NiPSysMeshEmitter update

Opens up access to the emitter-collection.

https://www.dropbox.com/s/hru8xv57hbxp7ya/NiPSysMeshEmitter.cpp.diff?dl=0
https://www.dropbox.com/s/sginxshn1de0l8d/NiPSysMeshEmitter.h.diff?dl=0

Python 3.x support

Update niflib to use Python3.x

hkPackedNiTriStripsData complement

Verify patch was added and if iterator is needed.

Provides access to the MOPP-info and corrects a few API texts.

https://www.dropbox.com/s/xpniuvaynwlwau2/hkPackedNiTriStripsData.h.diff?dl=0
https://www.dropbox.com/s/a90451eozrkelj0/hkPackedNiTriStripsData.cpp.diff?dl=0

How about this?
amorilia/niflib@50a17d6

The problem was that I needed access to , which you removed (the *Full() variant). That acess should be provided as well.

All parts of the hkTriangles are accessible - instead, I take it that you mean that the access isn't convenient for iteration, right? If you write the patch it will be even faster. :-)
Again, make sure setter and getter are symmetric, as in the remainder of the lib (it wasn't the case in your original patch).
Regarding the name, GetHkTriangles SetHkTriangles seems clearer?